SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Salary: EUR 80,000 - 100,000 / annual
Basalt is building AI agents powered by advanced machine learning models. As an ML Engineer, you will be responsible for the core technical work that makes these agents intelligent and reliable.
Your primary focus will be on building, training, and optimizing the large language models and neural components that power the product. You'll work across the full ML lifecycle: fine-tuning LLMs and embeddings models, designing and maintaining datasets, building preprocessing and training pipelines, running rigorous experiments to evaluate model behavior, and benchmarking different architectures to guide product decisions.
You'll use Python, PyTorch, HuggingFace, and modern ML tooling (Weights & Biases, etc.) to implement efficient training techniques including LoRA, QLoRA, distillation, quantization, and multi-GPU distributed training. You'll develop internal tooling to streamline training, evaluation, and data curation workflows, ensuring reproducibility and tracking improvements systematically.
You'll work closely with product and engineering teams to bring new ML capabilities into production, staying current with the latest advances in LLM training and evaluation methodologies. The role emphasizes a strong product mindset: you'll think about how models adapt to customer data, how they perform across diverse tasks, and how to scale training workflows efficiently.
Basalt values independent, rigorous, and proactive engineers who bring ideas, challenge assumptions, and elevate the team. The company is seeking pioneers motivated by AI challenges and excited about the possibilities of the technology.
The company offers a flexible remote policy with an expectation of regular office presence in Paris. Start date is flexible based on availability.
REQUIREMENTS:
- 3–5+ years of experience in Machine Learning, Deep Learning, or Applied AI
- Strong background in Python and PyTorch
- Hands-on experience fine-tuning LLMs or training neural models (including personal or research projects)
- Solid understanding of data preprocessing, training methodology, evaluation metrics, and experiment tracking
- Knowledge of efficient training techniques (LoRA, QLoRA, distillation, quantization, multi-GPU training)
- Familiarity with modern ML tooling (HuggingFace, Weights & Biases, etc.)
- Passion for building high-quality, reliable ML systems with a product mindset
- Independent, rigorous, and proactive approach to problem-solving
Note: The company encourages applications even if you don't meet every single requirement.
